Going to Mars and exploring the depths of space. Building micro-reactors that deliver limitless power. Designing novel therapeutics that eradicate diseases that have plagued humanity for millennia.

poolside founders Jason Warner and Eiso Kant see these as the three most significant leaps in human-led innovation that will be achieved during the 21st century. The pair operate under an Archimedean lens: solving AGI unlocks step-fold improvements in every other field, from space to climate to biology and beyond. But getting from here to there requires a new kind of software generation and reasoning company that trains models the way industry trains developers: sequentially.

Over the last 50 years, coding has accelerated human intelligence. We have, through building and architecting software systems, codified our reasoning into registers, caches, and logic gates. Teaching machines to reason is the final boss battle. poolside pursues a double mission: to first accelerate the billions of humans that today, cannot code, and through their experiences, build an AGI that can build anything.

Jumping In: the Cannonball and the Reverse Somersault

The best developer tools are not only effective, they are tasteful. Jason and Eiso have built their careers around an intimate understanding of the problems developers face.

While Jason served as GitHub's CTO, he incubated Copilot, Code Search and GitHub Actions. Before that, he led engineering at Heroku, which became the standard for web app development, and before that led engineering for Ubuntu Desktop, the most popular desktop Linux distro in the world. As early as 2018, Eiso was meanwhile building code search, code retrieval and ML-based code analysis tooling as the founder of source{d}.

Together, Jason and Eiso display a rare combination of tenacity, domain depth and product intuition. Their approach to talent is no different, hiring experts in GTM, design, product engineering and large model training right after the seed financing.

Meet the Founder: Dev-Focused, Through and Through

Jason has a long history of success leading teams building and shipping developer-focused products.

How It Works

As we’ve experienced in our own past lives as developers, getting any software project off the ground is a lesson in patience and coordination. Making sure the right versions of libraries are imported, making the right architectural decisions at the outset, navigating a large, convoluted file tree populated by colleagues that left the company years ago or scaffolding code that hasn’t been opened since the day it was generated.

What if we could just...build? It’s an almost heretical idea – to raise the level of abstraction for developers high enough that anyone can go from idea to running app in minutes, relying on an assistant that not only generates the code, but also helps the engineer make better decisions. As Jason and Eiso put it, the developer and AI mutually guide each other towards building the highest-value features in the most efficient way based on context present not just in code, but also in the cloud environment, IaC, application characteristics and more.

The question most enterprises have is: will it work on my applications? The product takes shape in four ways, depending on the specific problem the developer wants to solve.

1. Knowledge. As the developer prompts the assistant with what she needs to accomplish, the assistant returns the correct knowledge, whether it’s specific code blocks or deploy/config instructions. These are relevant not just for the code at hand, but also for the broader context of the application and its deployment configuration.
2. Action. If the developer opts in, the assistant will begin writing code, modifying configuration files and executing commands to build and test the code, acting as a junior engineer.
3. Architecture. Similarly, the assistant can execute on and drive architectural decisions that are expressed at a high level by the developer.
4. Application. Over time, developers will declare what apps or services they need built. The model will produce boilerplate and iteratively build specific functionality on top; allowing the world to program in natural language over code.

Ahead of the product launch, we saw a clear roadmap and product-market fit in our early conversations with Jason and Eiso.

An informed glimpse: Developers today must execute a tidal wave of manual tasks and write tedious glue code to keep their services and applications working. poolside envisions a near future where developers connect with technologies that help them translate natural language models and automated workflows into real-world developer operations – two or three levels of abstraction beyond a simple chat function to move from bugs and fixes to resolution and documentation.

poolside’s Assistant will translate ideas into applications over time, in brown and green field environments, with enough flexibility across the stack that developers like to use it. At present, the Assistant puts the power of poolside’s models directly inside developers’ IDE; operating as a Q&A interface that interacts with a developer as if it were a virtual teammate, as well as offering code completion features that leverage advanced context awareness to accurately provide real-time suggestions and predict developers' needs.

We’re believers in the speed and scale of poolside’s journey to general availability because of the proven track record Jason has earned through spearheading global product and eng teams, along with the foresight and startup leadership Eiso brings to the table.

What’s Next

Ultimately, poolside will automate every part of the development life cycle. In the process, poolside’s product will limit common headaches impacting developers’ ability to unblock, adjust or update code in real-time and at scale without disruptions.

“Compute is a proxy for understanding what's happening in the AGI race. If you don't have sufficient compute, you're just not in the race,” remarked Eiso. “With our latest round [of funding], we are one of very few companies in the world that have the level of resources to make meaningful progress.”

Living in the future has been a consistent thread tying Jason and Eiso’s career experiences together. Now, we’re excited to help them build it.